PTL 1 discloses an abnormality detection system using the relative comparison of power as an example of a solar battery abnormality detection system according to the related art.
In PTL 1, a sensor is connected in series between a plurality of solar battery arrays and a circuit which is called a power conditioner (PCS) and controls the output power of a solar battery and power generated by the solar battery is input to a solar battery array diagnosis device. The solar battery array diagnosis device measures the maximum power of each solar battery array for a plurality of days and creates a reference value. Power generated after the reference value is created is normalized by the reference value and is converted into weather information, such as sunny or cloudy. The weather information is compared in time series for each array and abnormality in the solar battery is detected on the basis of the comparison result. For example, when the solar battery operates without abnormality, the weather prediction results are identical to each other for all arrays. However, when abnormality occurs, the weather information items are not identical to each other in the comparison.
PTL 2 discloses a system capable of classifying the kinds of failure. First, a solar battery module which is presumed to be defective is separated from the power generation system and is connected to a failure detection system. The failure detection system operates a variable load in order to change the output voltage of the solar battery module from ‘0 V to an open voltage’ and performs curve tracing (measures the current-voltage characteristics). The measured current-voltage characteristics are converted into standard condition data, such as 25° C. and 1 kW/m2, on the basis of measurement information, such as the amount of solar radiation or the temperature, during the measurement of the solar battery.
In the system disclosed in PTL 2, the current-voltage characteristics when a plurality of failures occur are stored as reference characteristics in a memory in advance and the reference characteristics closest to the measured current-voltage characteristics are selected by, for example, a least-square method. Then, the kind of failure corresponding to the selected current-voltage characteristics is displayed on an LCD.